Monitoring of T&D Systems

Is there something useful in all the data?

Paul Myrda
Technical Executive

Stanford University
May 7, 2013
 Electric Power Research Institute

Generation Delivery Customer

Independent, Objective, Collaborative

© 2013 Electric Power Research Institute, Inc. All rights reserved. 2
 Today’s Power System

Generation Delivery Customer

Base Load Load Following Customer Interruptible

Generation
+ Generation
+/– Bulk Energy
Storage = Demand – Load DR

© 2013 Electric Power Research Institute, Inc. All rights reserved. 3

 Customer

Generation Delivery Customer

How will customers use electricity?

© 2013 Electric Power Research Institute, Inc. All rights reserved. 4
 Smart Grid Challenges

Top Down - 10 Smart Grid R&D Challenges

Standards & Interoperability
Communications Technology
Energy Mgmt Architecture & Integration
Security & Privacy
Renewable & DER Integration
Data Mgmt, Analysis & Visualization
Grid Management & Planning (Bulk)
Smart Grid Cost Benefit Analysis
Customer Integration Strategies
Advanced Technology Assessments
© 2013 Electric Power Research Institute, Inc. All rights reserved. 5
 Challenges: Turning Data into Opportunity & Value
Industry is creating more and more data

• From 1 Meter Read/mo to • No. Of PMU’s installed has

Hourly (720/mo) doubled in the past 12 months
• =71,900% increase in Data • Generating Terabytes of data
(((720-1)/1)*100%)
• Transitioning from
Implementation to Analysis

Management (Optimization) of the Grid = Management of Data

© 2013 Electric Power Research Institute, Inc. All rights reserved. 6
 Maximizing Data Value Throughout it’s Lifetime

Value of Individual Data Item Value of Data in Aggregate

(Real-Time Transactional Value) (Historical Analytical Value)

Complexity & Uncertainty

Data Value

Aggregate Data Early In

Life for More
Transactional Value

Age of Data
© 2013 Electric Power Research Institute, Inc. All rights reserved. 7
 The Opportunity

• Leverage Smart Grid Investments that are

Producing High-Quality Data Automated
Action
• Integrated Data Sources across different Decision
organizational groups Support

• Apply advanced data mining algorithms Information

• Demonstrate the data analytics applications
– Planning Operation, and Asset Raw Data
Management

Improve Reliability – Increase Operational Efficiencies

© 2013 Electric Power Research Institute, Inc. All rights reserved. 8

Data Integration for Asset Management and
Operations
 Backdrop

• Power Delivery has:

– Asset related data available in the field
– Knowledge about asset behavior
– Maintenance management systems & methods
– Sensor technology available to monitor assets

So what is the challenge?

© 2013 Electric Power Research Institute, Inc. All rights reserved. 10

 What if I only knew….

• When this circuit breaker last operated

– Was it slow or within specification
– Did it clear a fault or just opened for
maintenance
– Has it seen many faults, at what
energy level
– How are other similar units doing
• Is this transformer doing okay
– Is it gassing
– Are its winding loosening up
– Has it seen many faults, at what
energy level
– How are other similar units doing
© 2013 Electric Power Research Institute, Inc. All rights reserved. 11
 Our Generation Colleagues get it


(PRiSM Deployed Across Fossil Plants)

 (PRiSM Used For Gas Turbine Monitoring)


(Fleet Wide Generation Monitoring. Extending Transformer Monitoring)


(Fleet-wide Nuclear Implementation in Central Performance Monitoring Facility)

 (Fleet-wide Implementation With Over 2,000 models built for Gas and Coal
generation equipment)

Plant Process Computers Monitor Critical Elements

© 2013 Electric Power Research Institute, Inc. All rights reserved. 12

What’s the problem?
 How big is the gap?

Total Area is about

Byron 11,400 Sq. Mi.
Station 1300 Substations

Total Area is about

0.1 Sq. Mi.

Geography is over 100,000 times as big

© 2013 Electric Power Research Institute, Inc. All rights reserved. 14

 An Existing Architecture Example
Reporting & Visualization
Data Sources
Tools

Relational
Database
SCADA

Fault
SEL Files
Fault Files

Assets
STOMP

© 2013 Electric Power Research Institute, Inc. All rights reserved. 15

 An Example Future Architecture
Data Sources
TMDV Database

ETL Tools
SCADA

Fault
SEL Files
Fault Files

Assets
STOMP

TMDV Historian

© 2013 Electric Power Research Institute, Inc. All rights reserved. 16

 Uniqueness of Utility Data

High-frequency
Dynamic system Wind and solar
switching devices,
response
inverters output variation Day-ahead
(stability) T&D planning
Synchro-phasors Service scheduling
Protective relay Demand
response restoration
operations

seconds
10-6 10-3 10-0 103 106 109
millisecond second
minute hour day year decade
One AC Cycle Hour-ahead scheduling
and resolution of most Life span Life span
renewables integration studies of IT of OT
assets assets

The uniqueness of the utility market requires

unique technologies and architectures

© 2013 Electric Power Research Institute, Inc. All rights reserved. 17

 Other issues

• Data ownership
• Naming
• Location
• Understanding the broader enterprise value
• Data decimation
• Periodicity
• Latency
• Asset Testing
• System conditions
• System topology

© 2013 Electric Power Research Institute, Inc. All rights reserved. 18

 Field Data for Asset Management and
Operations
Problem: Utilities have numerous PI PI PI AF Enterprise
Process Book Performance
stand alone data sources that can Equations Network
Model
Visualization Asset Manager
be difficult to access and use Server Analytics (GIS and
History)

Enterprise Bus
A Solution: Use CIM to PI Historian
Asset Maintenance
Real-time
link and integrate Control Center Data Management Asset Test
Database Management
Historian System System
Visualization Real-time
Data
these data sources. Server Historian

Operations Bus
Integration Phase I –
Architecture Initial
Control 61850 Deployment
PDC Center Proxy
EMS Server in EPRI Lab
Substation
Substation Bus

RTU Relays
Substation
Gateway
PMUs Sensors

Using CIM models provides the basis for analytics that improves
decision making.
© 2013 Electric Power Research Institute, Inc. All rights reserved. 19
 Recent Project Demonstrations

Visualization

Approximately
250 ft radius
circle
Analytics

Approximately 50
ft radius circle
Integration

Application/ Application/ Application/ Application/ Application/

System System System System System
CCOTF Asset
Performance
Framework
© 2013 Electric Power Research Institute, Inc. All rights reserved. 20
 Synchrophasor-based Situational Awareness
and Decision Support
Online Monitoring Real-time Stability Assessment

Key Risk
Interface

Phase Clock on
Dominant oscillation mode the mode Simplified model on the interface Time

Phasor data
Scenario N
Scenario 2 Risk &
Scenario 1 Control
Timing
Look-Up Table

Control Action

Visualization at Control Room

Interconnected Power System
© 2013 Electric Power Research Institute, Inc. All rights reserved. 21
 Synchrophasor & Other Data for Disturbance Location
Identification
• Use data from PMUs and other
Generation Trip
sources to identify disturbance
location and magnitude
– frequency wave propagation
∆F and system frequency bias

© 2013 Electric Power Research Institute, Inc. All rights reserved. 22

 Synchrophasor-Based Early Warning of Inter-Area
Oscillations
100 • Identifying vulnerable grid
interfaces based on mode shapes
1-234
Angle Difference (deg.)

80 14-23

60 of inter-area oscillations
40
• Providing a risk index of angle
20
40 80 separation on the grid interfaces
120
Time (s)
160 200 240

180
Phase Difference (deg.)

1-234
14-23
90

0
40 80 120 160 200 240
Time (s)
1
1-234
0.8 14-23
Risk

0.6

0.4

0.2
40 80 120 160 200 240
Time (s)

© 2013 Electric Power Research Institute, Inc. All rights reserved. 23

 Synchrophasor & Other Data for System Dynamic Model
Development & Validation
Process Basics
Adequate system
• Use on-line disturbance data or staged disturbance
test data from generators, loads, et. al.
• Software tool to determine appropriate Extract DFR
data
generic dynamic model parameter
values Data post
processing
• Validate individual component models Disturbance Data
and overall system response
Model Derivation Baseline Model
Tool (PPPD) Structure & Data
Grid Disturbance
Model Derivation

PMU/DFR
at Plant
Compare derived
parameters w/
baseline data
Model Validation

Validate Power Plant Models for NERC MOD-26/27 Standards

© 2013 Electric Power Research Institute, Inc. All rights reserved. 24
 The Power Of Holistic Analytics

Online Dissolved
Gas in Oil
Analysis
Fluid Quality,
Moisture-
in-Oil

Transformer
Nameplate

Maintenance
History

Historical
Event Info

Readily Available EPRI Additional Condition

Sorting Process
Information Analysis Engine Information Indices

Number
of Units Effort # of Units Effort

© 2013 Electric Power Research Institute, Inc. All rights reserved. 25

 Using AMI Data to Identify Meter Phasing

Problem: Phasing errors are

common in GIS data.

A Solution: Use voltage and

current from substation SCADA
and customer AMI to estimate
phasing.

© 2013 Electric Power Research Institute, Inc. All rights reserved. 26

 Using AMI Data to Auto-Generate Secondary
Circuit Models
Problem: Utilities don’t have models
of secondaries, or they are poor.

A Solution: Use voltage and current

from AMI to auto-build secondary
circuit models.

Secondary model based on AMI data was better than the utility model

© 2013 Electric Power Research Institute, Inc. All rights reserved. 27

 Using AMI and Sensor Data for Distribution
State Estimation (DSE)
Problem: Utilities don’t have
way to identify and correct V
“bad” data V V

V
V

F
F F
V

A Solution: Use data from AMI F

F
F
and distribution sensors to
F F
determine current electrical
state of the distribution V

system

DSE improves reliability of advanced distribution applications

© 2013 Electric Power Research Institute, Inc. All rights reserved. 28

 Example: Combinations of Related Data Items

Problem: Utilities don’t have

way to evaluate efficiency
improvement benefits (e.g.
Voltage optimization)

A Solution: Combine data from

similar circuits and weather
data to identify benefits

© 2013 Electric Power Research Institute, Inc. All rights reserved. 29

 Pattern Recognition
Hydro Quebec: Using Feeder Monitors to
Locate Faults and Estimate Cause

A dangling phase conductor

that caused multiple
momentaries

Another case of a bad insulator

causing repeat momentaries

© 2013 Electric Power Research Institute, Inc. All rights reserved. 30

 Using AMI Data to Detect Stressed
Assets

Transformer load growth due to single

customer, 40 apartments, converted from gas
to electric space heaters - usage increased by
400%

Leveraging Data May Result in Avoiding Asset Failures

© 2013 Electric Power Research Institute, Inc. All rights reserved. 31

 Enabling Technologies that are Emerging

• Enterprise architecture
• Visualization technology
• Database technology
• Standards for
interoperability
• Sensor technology

© 2013 Electric Power Research Institute, Inc. All rights reserved. 32

Some Data Examples
 Relay Event Logs
(event, waveform, configuration)
• 030201,13384930,-4d,Magella,SOUTHSIDE LINE
PRI#16,APCo,,,04305n,07767w.SEL
FID=SEL-121G-5-R413-V656mptr12syzfs2-D941021-E2
Currents Voltages Relays Outputs Inputs
(amps) (kV)
52265L TCAAAAA DPBD5E
IPOL IR IA IB IC VA VB VC 01171O PL1234L TTTC2T
8 15 -8 -102 102 -63.4 57.0 6.8 ...... ....... ....*.
0 -15 94 -64 -26 -28.8 -40.6 69.6 ...... ....... ....*.
-8 -15 4 98 -106 63.5 -57.0 -6.8 ...... ....... ....*.
0 30 -98 64 30 28.8 40.7 -69.6 ...... ....... ....*.
8 -15 0 -94 106 -63.5 56.9 6.9 ...... ....... ....*.
0 -15 98 -64 -30 -28.7 -40.8 69.6 ...... ....... ....*.
-8 15 0 94 -102 63.5 -56.9 -7.0 ...... ....... ....*.
0 0 -98 68 26 28.6 40.9 -69.6 ...... ....... ....*.

Event : EXT Location : mi ohms sec

Duration: Flt Current:
R1 =0.40 X1 =2.26 R0 =1.96 X0 =6.90 LL =3.01
CTR =240.00 PTR =1000.00 MTA =75.00 LOCAT=Y
79OI1=10000.00 79OI2=10000.00 79OI3=10000.00 79RS =8000.00
Z1% =90.00 Z2% =140.00 Z3% =171.00
Z2SP =0.00 Z2DP =25.00 Z3DP =90.00

© 2013 Electric Power Research Institute, Inc. All rights reserved. 34

 Kohonen Neural Networks

© 2013 Electric Power Research Institute, Inc. All rights reserved. 35

 Relay Signature Analysis
(cluster analysis & movie)

© 2013 Electric Power Research Institute, Inc. All rights reserved. 36

 Thank you!

Paul Myrda
pmyrda@epri.com
708-479-5543

Together . . . Shaping the Future of Electricity